Agonist mediated internalization of M2 mAChR is β-arrestin-dependent

Background Muscarinic acetylcholine receptors (mAChRs) undergo agonist-promoted internalization, but evidence suggesting that the mechanism of internalization is β-arrestin dependent has been contradictory and unclear. Previous studies using heterologous over-expression of wild type or dominant-negative forms of β-arrestins have reported that agonist-promoted internalization of M2 mAChRs is a β-arrestin- and clathrin-independent phenomenon. In order to circumvent the complications associated with the presence of endogenous β-arrestin that may have existed in these earlier studies, we examined agonist-promoted internalization of the M2 mAChR in mouse embryonic fibroblasts (MEFs) derived from β-arrestin knockout mice that lack expression of either one or both isoforms of β-arrestin (β-arrestin 1 and 2). Results In wild type MEF cells transiently expressing M2 mAChRs, 40% of surface M2 mAChRs underwent internalization and sorted into intracellular compartments following agonist stimulation. In contrast, M2 mAChRs failed to undergo internalization and sorting into intracellular compartments in MEF β-arrestin double knockout cells following agonist stimulation. In double knockout cells, expression of either β-arrestin 1 or 2 isoforms resulted in rescue of agonist-promoted internalization. Stimulation of M2 mAChRs led to a stable co-localization with GFP-tagged β-arrestin within endocytic structures in multiple cell lines; the compartment to which β-arrestin localized was determined to be the early endosome. Agonist-promoted internalization of M2 mAChRs was moderately rescued in MEF β-arrestin 1 and 2 double knockout cells expressing exogenous arrestin mutants that were selectively defective in interactions with clathrin (β-arrestin 2 ΔLIELD), AP-2 (β-arrestin 2-F391A), or both clathrin/AP-2. Expression of a truncated carboxy-terminal region of β-arrestin 1 (319–418) completely abrogated agonist-promoted internalization of M2 mAChRs in wild type MEF cells. Conclusion In summary, this study demonstrates that agonist-promoted internalization of M2 mAChRs is β-arrestin- and clathrin-dependent, and that the receptor stably co-localizes with β-arrestin in early endosomal vesicles.


Background
Muscarinic acetylcholine receptors belong to the superfamily of G-protein coupled receptors (GPCRs) that are commonly expressed in a variety of tissues and are classified into five known subtypes (M 1 -M 5 mAChR). M 1 , M 3 , and M 5 mAChRs are selectively coupled to G q proteins while M 2 and M 4 mAChRs are linked to G i /G 0 proteins [1,2]. M 2 mAChRs are the primary muscarinic subtype in the heart where their stimulation leads to the regulation of myocardial contractility [3]. As with other GPCRs, M 2 mAChR activity is tightly regulated by desensitization and internalization. These regulatory mechanisms are typically associated with receptor phosphorylation followed by either recycling or down-regulation [4][5][6][7][8][9].
In addition to desensitization and internalization, βarrestins are known to play a role in other cellular processes that include intracellular trafficking and signalling [12]. Association of β-arrestin with agonist-occupied receptors has been shown to initiate intracellular signalling by functioning as an assembly site for signalling components such as Src, JNK3, and ERK1/2 [14][15][16][17]. Therefore, β-arrestin-receptor complexes can lead to cytosolic retention and activation of signalling molecules following receptor-mediated signalling at the cell surface. The physiological roles of this process include decreasing cell proliferation and regulating cytoskeletal rearrangements by spatially restricting ERK activation to the cytosol [16,18]. Recent reports have also suggested that βarrestins can function at post-endocytic stages to regulate receptor sorting. It has been shown that receptors exhibit differential affinities for β-arrestin and therefore are classified into two groups [19]. Class A receptors (including β 2 AR and dopamine receptors) are thought to interact with β-arrestin at the plasma membrane but immediately disassociate following localization to clathrin-coated pits. Hence receptors enter early endosomes devoid of β-arrestin and are typically resensitized and rapidly recycled [20]. In contrast, Class B receptors (vasopressin-V 2 R, angiotensin-AT 1A R, and neurotensin receptors) stably associate with β-arrestin so that β-arrestin/receptor complexes remain intact and are internalized into juxtanuclear endosomal compartments [21]. This interaction can persist for prolonged periods of time. This stable association may dictate the kinetics of receptor recycling since AT 1A R and V 2 R recycle very slowly [20,21]. A functional consequence of β-arrestin association may also be to facilitate receptor down-regulation.
The role of β-arrestins in regulating the trafficking of M 2 mAChRs has been contradictory and unclear. Reports have demonstrated that phosphorylation by GRK2 on serine/threonine residues in the third intracellular loop of M 2 mAChRs recruits β-arrestin and leads to receptor desensitization and subsequent internalization [7].
Whether β-arrestin is involved directly in agonist-promoted endocytosis of M 2 mAChRs remains unclear. Indeed over-expression of β-arrestin has been reported to increase agonist-promoted internalization of M 2 mAChRs but not of M 1 or M 3 mAChRs [22]. Furthermore, Claing et al. have shown that M 2 mAChRs internalize in a dynaminand β-arrestin-insensitive manner when expressed in HEK293 cells [23]. Others have reported that the Arf6 GTPase (ADP-ribosylation factor 6) facilitates M 2 mAChR entry into primary vesicles, which fuse with clathrinderived early endosomes [24,25]. These data do not necessarily rule out β-arrestin as a regulator in agonist-promoted endocytosis of M 2 mAChRs. Therefore, to clarify whether agonist-promoted internalization of M 2 mAChRs is arrestin dependent, we utilized mouse embryonic fibroblasts (MEFs) derived from β-arrestin null mice that lack expression of one or both isoforms (β-arrestin 1 and 2) and their wild type littermates as control cells [26]. Here we report that agonist-promoted internalization of M 2 mAChRs is β-arrestin dependent and M 2 mAChRs form stable complexes with β-arrestin at the early endosome. Furthermore, we demonstrate that agonist-promoted internalization of M 2 mAChRs is clathrindependent. These results suggest that β-arrestin plays an important role in regulating M 2 mAChR activity.

Results
To determine whether the MEF cells used in this study expressed endogenous mAChRs, we performed RT-PCR aimed at detecting mRNA encoding M 1 , M 2 and M 4 mAChR subtypes. As positive controls, we used postnatal rat cerebellum tissue for M 2 mAChR mRNA and postnatal rat cortical tissue for M 1 and M 4 mAChR mRNA. RT-PCR analysis clearly demonstrated that MEF wild type as well as MEF double knockout cells (MEF KO1/2) did not express mRNA encoding M 1 , M 2 , or M 4 mAChR subtypes (Fig. 1). Accordingly, radioligand-binding assays also confirmed that MEF wild type as well as MEF KO1/2 did not express mAChRs at any detectable level (data not shown). Therefore, we concluded that MEF cells do not express endogenous mAChRs.
To examine whether ectopically expressed M 2 mAChRs undergo agonist-promoted internalization in MEFs, we transiently transfected MEF wild type and corresponding β-arrestin null cells with a plasmid encoding a FLAGtagged porcine M 2 mAChR. Following 24 h transfection, MEF wild type, MEF KO1, MEF KO2, and MEF KO1/2 cells were stimulated with 1 mM carbachol for 1 h at 37°C. The number of receptors remaining at the cell surface was measured using a saturating concentration of the hydrophilic ligand [ 3 H]-NMS. Approximately 40% of surface M 2 mAChRs were internalized in wild type MEF cells while M 2 mAChRs in MEF KO1 and MEF KO2 cells were internalized by 33% and 42%, respectively. In contrast, M 2 mAChRs were not internalized in MEF KO1/2 ( Fig.  2A). These results demonstrated that exogenously expressed M 2 mAChRs undergo agonist-promoted internalization in MEF wild type cells and either β-arrestin isoform was sufficient for sequestration. To further evaluate where M 2 mAChRs were localized, we used confocal immunofluorescence microscopy in MEF wild type or MEF KO1/2 cells transiently expressing a FLAG-tagged M 2 mAChR in the absence or presence of carbachol. As indicated in Figure 2B, diffuse cell surface localization of M 2 mAChRs was observed prior to carbachol addition in both MEF phenotypes. Upon addition of agonist, M 2 mAChRs in MEF wild type cells redistributed into discrete intracellular vesicles dispersed throughout the cell while M 2 mAChRs expressed in MEF KO1/2 cells remained at the cell surface (Fig. 2B). The diffuse pattern shown in MEF KO1/2 cells represents surface plasma membrane localization since the absence of detergent leads to an identical staining pattern as seen in untreated cells (data not shown).
Mouse embryonic fibroblasts (MEF) cells do not express mRNA encoding M 1 , M 2 or M 4 mAChR subtypes The FLAG-tag is located at the N-terminus of the receptor and is accessible to exogenously added antibody even in the absence of detergent.
To determine whether selectivity existed between β-arrestin isoforms in their ability to mediate agonist-promoted internalization of M 2 mAChRs, we examined agonist pro-Agonist-promoted internalization of M 2 mAChR in MEFs is β-arrestin-dependent Figure 2 Agonist-promoted internalization of M 2 mAChR in MEFs is β-arrestin-dependent. A.) Approximately 24 h following transfection with M 2 mAChR, cells were stimulated with 1 mM carbachol for 1 h and agonist-promoted internalization was determined using [ 3 H-NMS]. Data are presented as the mean ± standard error from 3 separate experiments with each experiment consisting of 8 to 11 independent determinations. Statistical test was performed using ANOVA with the post hoc Bonferroni/Dunn test (asterisk indicates * p < 0.001). B.) Cells were transfected as described above and then incubated in the presence or absence of 1 mM carbachol for 30 minutes prior to indirect immunofluorescence as described in Methods. Images were acquired at 40X. moted internalization in MEF KO1/2 cells co-expressing M 2 mAChR and FLAG-tagged β-arrestin 1 and/or 2 ( Fig.  3A). Western blotting analysis confirmed that FLAGtagged β-arrestins were expressed (Fig. 3B). Cells were treated with 1 mM carbachol for 1 h and the extent of receptor internalization was assessed using [ 3 H]-NMS. MEF KO1/2 cells reintroduced with β-arrestin 1, β-arrestin 2, or both isoforms exhibited M 2 mAChR uptake similarly (Fig. 3A). These data suggest that not only is agonist-promoted internalization of M 2 mAChR β-arrestin-dependent but also there is no selectivity between β-arrestin isoforms ( Fig. 2A and 3A). To assess whether stimulated and internalized M 2 mAChRs co-localize with β-arrestin 1 or 2, we reintroduced GFP-tagged β-arrestin 1, 2, or both isoforms with FLAG-tagged M 2 mAChRs into MEF KO1/2 cells and assessed their localization by immunofluorescence microscopy. Internalized M 2 mAChRs remained associated with β-arrestin 1-GFP (data not shown) or β-arrestin 2-GFP ( Fig.  4) in intracellular compartments following 30 minutes stimulation with 1 mM carbachol. To determine if this phenomenon occurs in other cell types we expressed M 2 mAChRs in HeLa, COS-7, and rat aortic smooth muscle cells (RASMCs). As observed in MEF KO1/2 cells, internalized M 2 mAChRs remained co-localized with β-arrestin 2-GFP in HeLa, COS-7, and RASMCs (Fig. 5). These results demonstrate that agonist-promoted internalization of the M 2 mAChR is β-arrestin-dependent, and that internalized M 2 mAChRs stably associate with either β-arrestin isoform in multiple cell lines.
Recent studies by Santini and co-workers [30] showed that agonist-mediated activation of the β 2 AR was still capable of inducing recruitment into clathrin coated pits in cells expressing mutant arrestin proteins that were defective in binding with clathrin or AP-2, albeit to a reduced degree. Expression of the truncated COOH-terminal region of β-arrestin 1 (319-418), which contains a clathrin binding site but lacks receptor binding, completely inhibited the β 2 AR mediated clustering of clathrin coated pits [31]. Therefore, we conducted experiments with the truncated β-arrestin 1 (319-418) to determine whether agonist-promoted internalization of the M 2 mAChR in MEFs would be affected. Transient expression of the truncated β-arrestin 1 (319-418) completely inhibited the agonist-promoted internalization of the M 2 mAChR in MEF wild type cells (Fig. 6B). Thus, it could be argued that the agonist-promoted internalization of M 2 mAChR involved a clathrin-dependent pathway. However, as shown previously, expression of an arrestin 2 mutant that was defective in interaction with both clathrin and AP-2 only moderately antagonized the agonistpromoted internalization of M 2 mAChR in MEF KO1/2 cells (Fig. 6A). It could be argued that this arrestin mutant, defective in clathrin/AP-2 binding, was still capable of interacting with clathrin/AP-2, albeit to a significantly reduced degree. Thus, it is reasonable to conclude that the agonist-promoted internalization of M 2 mAChRs was clathrin-dependent. Based upon the findings described above, we sought the identity of the endosomal structures to which β-arrestin localized following M 2 mAChR activation. We performed co-localization analyses using markers of the early endosome, the early endosomal autoantigen-1 (EEA-1) and the transferrin receptor (TfnR), in combination with β-arrestin 1-GFP. β-arrestin 1-GFP and FLAG-M 2 mAChRs were co-expressed in HeLa cells, and cells were stimulated with 1 mM carbachol for 30 minutes. Our results showed that β-arrestin 1-GFP significantly co-localized with EEA-1 and TfnR (as indicated by arrows in Fig. 7). β-arrestin 1-GFP was not observed to be associated with EEA-1 or TfnR in unstimulated HeLa cells (Fig. 7). These results indicate that once M 2 mAChRs are internalized via a β-arrestin dependent pathway, they remain co-localized with βarrestin in clathrin-derived early endosomes. To address whether other muscarinic receptor subtypes stably associate with β-arrestin in endosomes, we co-expressed HAtagged M 1, M 3, M 4, and M 5 mAChRs with β-arrestin 2-GFP in MEF wild type cells and assessed β-arrestin localization using confocal microscopy (Fig. 8). The upper right inset in each frame of the figure shows the localization of the muscarinic receptor subtype for a small section of the cell co-expressing β-arrestin 2-GFP. Overlay images indicate co-immunostaining of mAChRs (red) with β-arrestin 2-GFP (green) and their extent of co-localization (yellow). In the absence of carbachol, β-arrestin 2-GFP was diffusely localized in the cytosol of cells expressing M 1 -M 5 mAChR subtypes (Fig. 8, 0 min). Following 30 minute carbachol stimulation only cells expressing human FLAGtagged M 2 mAChRs exhibited β-arrestin 2-GFP localiza-tion in intracellular compartments as shown by arrows indicating overlap and corresponding overlay image (Fig.  8, 30 min); in cells expressing other receptor subtypes, βarrestin 2-GFP remained diffusely distributed. Hence, only cells expressing the FLAG-tagged M 2 mAChR subtype exhibited a stable interaction with β-arrestin at intracellular sites compared to the other muscarinic subtypes.

Discussion
In the present study, we investigated the role of β-arrestin in agonist-promoted internalization of the M 2 mAChR, which has previously been reported to be β-arrestin independent. In previous studies, heterologous over-expression of wild type and dominant-negative forms of arrestins was used to assess the function of these proteins Expression of β-arrestin 1 or 2 rescued agonist-promoted internalization of M 2 mAChRs in MEF KO1/2 cells [ 22,32]. Unfortunately, such studies are difficult to interpret because of the complications associated with endogenous proteins. In an attempt to alleviate these complications, we utilized mouse embryonic fibroblasts (MEFs) derived from β-arrestin knockouts in which endogenously expressed β-arrestin 1 and 2 have been genetically eliminated [26]. These cells provide us a unique opportunity to assess whether β-arrestin proteins are involved in the process of agonist-promoted internalization of M 2 mAChRs. Herein, we show that agonist-promoted endocytosis of the M 2 mAChR is β-arrestinand clathrin-dependent.
Expression of β-arrestin mutants deficient in clathrin and/or AP-2 binding interaction partially supports agonist-promoted internalization of M 2 mAChRs in MEF K/O1/2, while expression of truncated carboxyl-terminal region of β-arrestin 1 (319-418) completely blocked agonist promoted M 2 mAChR internalization in MEFwt cells Addition of agonist leads to the redistribution of β-arrestin 1-GFP to early endosomal structures in the cytosol Figure 7 Addition of agonist leads to the redistribution of β-arrestin 1-GFP to early endosomal structures in the cytosol. HeLa cells were transiently transfected with human FLAG-tagged M 2 mAChR and β-arrestin 1-GFP and treated with 1 mM carbachol for 30 minutes. Cells were processed for confocal microscopy. β-arrestin 1-GFP complexes localized to the early endosome as shown by colocalization with markers of that compartment (EEA-1 and TfnR). Arrows indicate signficant overlap between TfnR or EEA-1 with β-arrestin 1-GFP. Confocal images are representative of three independent experiments.
Both β-arrestin 1 and 2 isoforms were reported to form high affinity complexes with the agonist-activated M 2 mAChR [33], suggesting that either isoform is capable of mediating agonist-promoted internalization of the receptor. In corroboration with these findings, we observed no selectivity between β-arrestin isoforms in mediating agonist-promoted internalization of M 2 mAChRs. Perhaps, this lack of selectivity between β-arrestin 1 and 2 may explain why using over-expression of a single mutant form of β-arrestin fails to completely block the agonistpromoted internalization of M 2 mAChRs. Interestingly, our studies further revealed that β-arrestin remained stably associated with the M 2 mAChR in juxtanuclear endosomes for prolonged periods of time following agonist exposure. Given that MEF cells do not endogenously express mAChRs, we compared our observations in a physiologically relevant cell line (RASMCs) and two model cell lines (HeLa and COS-7). Similar findings were also observed in these cells. Closer examination of β-arrestin post-endocytic trafficking revealed that M 2 mAChR stimulation led to arrestin redistribution into Tfn and EEA-1 positive compartments, markers of the early endosome. In accordance with our findings, Delaney et al. have reported that stimulated M 2 mAChRs internalize in a manner that quickly merges with clathrin-derived early endosomes [25]. M 2 mAChRs follow the general pattern utilized by most GPCRs in that they are internalized via a β-arrestindependent mechanism. Additionally, the stable binding of β-arrestin with activated M 2 mAChRs within microcompartments follows the paradigm of other Class B GPCRs. Implications of these findings are that β-arrestin may dictate the intracellular trafficking and/or signalling of the M 2 mAchRs. Since β-arrestin has emerged as a versatile adaptor and scaffolding protein, its role in regulating M 2 mAChR-dependent cellular activity may be significant. It has been shown that β-arrestins interact with trafficking machinery such as Arf6, RhoA, NSF, and a variety of signalling proteins such as ASK1, JNK3, and ERK1/2 [34]. Stable β-arrestin/receptor complexes, as exhibited by Class B receptors, appear to redirect signalling complexes to the cytoplasm thereby activating cytoplasmic targets while preventing ERK translocation to the nucleus [15,16,35]. The physiological role of this process may be to participate in actin cytoskeleton reorganization and chemotaxis [18,36]. With regard to intracellular trafficking, patterns of β-arrestin binding to activate receptors appear to modulate receptor recycling and/or degradation [37]. Class A receptors are typically resensitized and subsequently recycled while Class B receptors undergo slow recycling and/or down-regulation. M 2 mAChRs have been shown to undergo slow recycling back to the plasma membrane upon agonist removal [38]. What role or roles β-arrestin plays in M 2 mAChR recycling and/or degrada-tion is currently unknown. The functional consequence of stable β-arrestin/M 2 mAChR complexes remains to be determined.
Previous studies have suggested that M 2 mAChR internalization does not proceed through a β-arrestin/clathrin mediated pathway [22,23,28]. For example, Delaney and co-workers [25] previously reported that M 2 mAChRs internalized by a clathrin-independent pathway based upon the use of a dominant-negative K44A dynamin-1 mutant. However, expression of a N-terminal deletion dynamin-1 mutant N272 that lacks the complete GTPbinding domain, unlike K44A dynamin, strongly inhibited agonist-promoted M 2 mAChR internalization [39]. Therefore, we conducted experiments with arrestin mutants that were selectively deficient in interaction with clathrin, AP-2, or both clathrin and AP-2, to determine whether agonist mediated internalization of M 2 mAChRs was clathrin-dependent. Expression of arrestin mutants defective in interaction with either clathrin (β-arrestin 2-ΔLIELD) or AP-2 (β-arrestin 2-F391A) failed to antagonize M 2 mAChR internalization. Moreover, over-expression of a dominant-negative arrestin mutant that was defective in interaction with both clathrin and AP-2 only modestly antagonized M 2 mAChR internalization in MEF KO1/2 cells. Thus, it is reasonable to conclude that these data corroborate previous studies indicating that M 2 mAChR internalization is clathrin-independent. However, Santini and co-workers [30] have reported that arrestin mutants with impaired binding to clathrin or AP-2 were still capable of displaying recruitment of β 2 AR to clathrin-coated pits, albeit to a reduced degree. Therefore, it may be premature to conclude that M 2 mAChR internalization is βarrestin-dependent but clathrin/AP-2-independent. Expression of the truncated carboxy-terminal region of βarrestin 1, which contained the clathrin interaction site, has been shown to completely abrogate β 2 AR mediated clustering of clathrin coated pits [31]. Exogenous expression of this mutant completely block agonist-promoted internalization of M 2 mAChRs in wild type MEFs. Collectively, these results indicate that agonist-promoted internalization of M 2 mAChRs is β-arrestin-dependent and most likely clathrin/AP-2-dependent. However, we cannot rule out that the C-terminal region of arrestin 1 is interacting with another factor, independent from clathrin/AP-2 that may be responsible for mediating internalization of the M 2 mAChR. Indeed previous studies have shown that the Arf6 GTPase regulates agonist-promoted endocytosis of the M 2 mAChR [24,25]. It has been shown that β 2 AR stimulation leads to activation of Arf6 GTPase, which facilitates receptor endocytosis [40]. It is feasible that sequestration of M 2 mAChR requires activation of Arf6 GTPase by a β-arrestin-mediated pathway, which may be an important component of agonist-promoted internalization of the M 2 mAChR. This would corroborate previous studies, which indicate a critical role for Arf6 GTPase in mediating agonist-promoted M 2 mAChR internalization [24].
The differential trafficking of β-arrestin with mAChRs to endosomes appears to be subtype specific. There are five muscarinic subtypes termed M 1 mAChR-M 5 mAChR. M 1 , M 3, and M 5 mAChRs couple to G q proteins and activate phospholipase C whereas M 2 mAChR and M 4 mAChR couple to G i/o to inhibit adenylyl cyclase and activate K + channels [1,2]. As shown in Figure 8, stimulated muscarinic subtypes aside from M 2 mAChRs are sequestered into endocytic vesicles that are devoid of β-arrestin. It has been shown that M 1 mAChR, M 3 mAChR, and M 4 mAChR require β-arrestin in mediating agonist-promoted internalization [23] so we do not rule out the possibility that arrestin is recruited to the plasma membrane following stimulation and then rapidly disassociates from the receptor. It is possible that carbachol may induce receptor conformations that may not promote stable β-arrestin associations with the other mAChR subtypes. However, sequence alignment of the M 2 and M 4 mAChR (using the T-coffee program) revealed that the subtypes exhibit high sequence similarities; interestingly, the sequence differences lie in the third intracellular loop, specifically at residues 293-313 within the M 2 mAChR. As described by Pals-Rylaarsdam and others, a cluster of serine and threonine sites at positions 307-311 undergo agonist promoted phosphorylation, which is necessary and sufficient for β-arrestin interaction [6]. This site may be important for designating stable interactions with β-arrestin. M 2 mAChR sequences downstream from this site at 348-368 also differ significantly from the M 4 mAChR suggesting that an additional motif may be involved.

Conclusion
In summary, the data presented in this study demonstrate that the agonist-promoted endocytosis of the M 2 mAChR subtype occurs via an arrestin dependent pathway in MEF cells. Exogenously expressed β-arrestin proteins remained stably associated with the M 2 mAChR upon entry into early endosomal compartments. The lack of stable β-arrestin interaction with other mAChR subtypes suggests a unique role of β-arrestin in regulating activity of the M 2 mAChR subtype.  [30,31]. The MEF wild type, β-arrestin 1 and 2 single knockouts, βarrestin 1 and 2 double knockout cells, and constructs for FLAG-tagged β-arrestin 1 and 2 were kindly provided by Dr. Robert Lefkowitz (Duke University Medical Center) [26]. Constructs encoding β-arrestin 2-GFP and β-arrestin 1-GFP were generous gifts from Dr. Stefano Marullo and have been previously described [41].

Cell Culture and Transient Transfection
HeLa, MEF wild-type, MEF single and double β-arrestin knockout, RASMCs, and COS-7 cells were maintained in DMEM supplemented with 10% fetal bovine serum (FBS), 100 I.U./ml penicillin, and 100 μg/ml streptomycin at 37°C with 5% CO 2 . For immunocytochemistry, HeLa cells were grown on glass coverslips at a density of 120,000 cells/well in six-well dishes and transfected with EX-GEN or LipofectAMINE 2000 according to the manufacturer's protocol using 1 μg of DNA/well. For ligand binding assays, MEF cells were plated at 80,000 cells/well in 24 well plates and transfected with EX-GEN or Lipo-fectAMINE 2000 according to the manufacturer's protocol using 1 μg of DNA/well.

Radioligand Binding Assay
Receptor internalization was determined by measuring the binding of the membrane impermeable muscarinic antagonist [ 3 H]-N-methylscopolamine ([ 3 H]-NMS) to intact cells as previously described [42]. Briefly, 24-42 h after transfection, MEF cells cultured in 24-well plates were treated or not treated with 1 mM carbachol for 60 min at 37°C. Cultures were washed twice with 1 ml of icecold PBS, and labelled with 720 fmol of [ 3 H]-NMS in 1 ml PBS for 4 h at 4°C. Non-specific binding was determined as the bound radioactivity in the presence of 1 μM atropine. Labelled cells were washed two times with 1 ml of ice-cold PBS, solubilized in 0.5 ml of 1% Triton X-100 and combined with 3.5 ml of scintillation fluid followed by Internalized M 2 mAChRs exhibit a differential affinity for β-arrestin 2-GFP compared to other muscarinic subtypes Figure 8 Internalized M 2 mAChRs exhibit a differential affinity for β-arrestin 2-GFP compared to other muscarinic subtypes. HeLa cells were transiently co-transfected with plasmids encoding β-arrestin 2-GFP and either HA-tagged M 1 , M 3, M 4, M 5 mAChR or FLAG-tagged M 2 mAChR. Cells were untreated or treated with 1 mM carbachol for 0 min or 30 min. Grayscale image indicates β-arrestin 2-GFP localization while the upper right inset indicates immunostaining of the mAChR in a small section of the cell (outline). Arrows indicate overlap between internalized M 2 mAChRs and β-arrestin 2-GFP. Overlay represents co-immunostaining of mAChR (red) and β-arrestin 2-GFP expression (green) and their colocalization (yellow).